## Thank you for using fastDummies!
## To acknowledge our work, please cite the package:
## Kaplan, J. & Schlegel, B. (2023). fastDummies: Fast Creation of Dummy (Binary) Columns and Rows from Categorical Variables. Version 1.7.1. URL: https://github.com/jacobkap/fastDummies, https://jacobkap.github.io/fastDummies/.
## Loading required package: Matrix
##
## Attaching package: 'readr'
## The following object is masked from 'package:scales':
##
## col_factor
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
##
## Attaching package: 'gridExtra'
## The following object is masked from 'package:dplyr':
##
## combine
## -- Attaching core tidyverse packages ------------------------ tidyverse 2.0.0 --
## v forcats 1.0.0 v stringr 1.5.1
## v lubridate 1.9.3 v tibble 3.2.1
## v purrr 1.0.2 v tidyr 1.3.1
## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x readr::col_factor() masks scales::col_factor()
## x gridExtra::combine() masks dplyr::combine()
## x purrr::discard() masks scales::discard()
## x tidyr::expand() masks Matrix::expand()
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()
## x tidyr::pack() masks Matrix::pack()
## x tidyr::unpack() masks Matrix::unpack()
## i Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
1. Load data, preprocess
Implementation for DiD, SC, Synth. DiD, Generate plots for different
outcome variables
# Get vector of cohort years
cohorts <- sort(unique(dat$treatment_year))
cohorts <- cohorts[!is.infinite(cohorts)]
# Initialize a list to store results
results <- list()
#' Function that implements DiD, SC, and Synth. DiD and plots outcomes
#' @param data panel data with state-year as units
#' @param outcome_var dependent variable we want to measure
#' @param cohorts vector of treatment years present throughout the data
run_synthdid <- function(data, outcome_var, cohorts) {
outcome_labels <- c(
pct_in_comp_mco = "Percent in Comp. MCO",
total_medicaid = "Total Medicaid Spending",
log_total_spend = "Log(Total Medicaid Spending)",
all_medicaid_spending_per_cap = "Total Medicaid Spending per Capita",
log_all_medicaid_spending_per_cap = "Log(Total Medicaid Spending per Capita)"
)
outcome_label <- outcome_labels[[outcome_var]]
# Loop through each cohort
for (cohort_year in cohorts) {
tryCatch({
# Subset data for the current cohort
cohort_data <- data %>%
filter(treatment_year == cohort_year | is.infinite(treatment_year))
# Reshape the cohort data to wide format
Y <- cohort_data %>%
select(state, year, !!sym(outcome_var)) %>%
pivot_wider(names_from = year,
values_from = !!sym(outcome_var)) %>%
column_to_rownames("state") %>%
as.matrix()
# Extract treatment timing for the cohort
treatment_timing <- cohort_data %>%
group_by(state) %>%
summarize(treatment_year = unique(treatment_year)) %>%
pull(treatment_year)
# Number of control units (N0)
N0 <- sum(treatment_timing == Inf)
# Number of pre-treatment periods (T0)
T0 <- cohort_year - min(cohort_data$year)
# Check if Y has valid dimensions
if (nrow(Y) == 0 || ncol(Y) == 0) {
stop("Y has invalid dimensions. Skipping cohort year ", cohort_year)
}
# Estimate the treatment effect using synthdid
tau.hat <- synthdid_estimate(Y, N0, T0)
# Define estimators
estimators <- list(did = did_estimate,
sc = sc_estimate,
sdid = synthdid_estimate)
# Apply estimators
estimates <- lapply(estimators, function(estimator) {estimator(Y, N0, T0)})
# Generate the plot
plot <- synthdid_plot(estimates[1:3],
control.name = 'control',
treated.name = "treated", # Use the treated unit's name
facet.vertical = FALSE,
lambda.comparable = TRUE,
se.method = 'none',
trajectory.linetype = 1,
line.width = .75,
effect.curvature = -.4,
trajectory.alpha = .7,
effect.alpha = .7,
diagram.alpha = 1,
onset.alpha = .7) +
labs(title = paste0("Estimators - ", outcome_label),
subtitle = paste0("Treatment Year = ", cohort_year),
x = "Year",
y = outcome_var)
print(plot)
# Print a success message
message("Plot saved for cohort year ", cohort_year)
}, error = function(e) {
# Print an error message and skip the cohort year
message("Error in cohort year ", cohort_year, ": ", e$message)
})
}
}
Percent in Comprehensive Managed Care
run_synthdid(data, "pct_in_comp_mco", cohorts)
## Error in cohort year 1992: dim(X) must have a positive length
## Plot saved for cohort year 1994
## Plot saved for cohort year 1995
## Plot saved for cohort year 1996
## Plot saved for cohort year 1997
## Plot saved for cohort year 1998
## Plot saved for cohort year 2006
## Plot saved for cohort year 2007
## Plot saved for cohort year 2011
## Plot saved for cohort year 2012
## Plot saved for cohort year 2014
## Plot saved for cohort year 2016
## Plot saved for cohort year 2021
Total Medicaid Spending
run_synthdid(data, "total_medicaid", cohorts)
## Error in cohort year 1992: dim(X) must have a positive length
## Plot saved for cohort year 1994
## Plot saved for cohort year 1995
## Plot saved for cohort year 1996
## Plot saved for cohort year 1997
## Plot saved for cohort year 1998
## Plot saved for cohort year 2006
## Plot saved for cohort year 2007
## Plot saved for cohort year 2011
## Plot saved for cohort year 2012
## Plot saved for cohort year 2014
## Plot saved for cohort year 2016
## Plot saved for cohort year 2021
Log(Total Medicaid Spending)
run_synthdid(data, "log_total_spend", cohorts)
## Error in cohort year 1992: dim(X) must have a positive length
## Plot saved for cohort year 1994
## Plot saved for cohort year 1995
## Plot saved for cohort year 1996
## Plot saved for cohort year 1997
## Plot saved for cohort year 1998
## Plot saved for cohort year 2006
## Plot saved for cohort year 2007
## Plot saved for cohort year 2011
## Plot saved for cohort year 2012
## Plot saved for cohort year 2014
## Plot saved for cohort year 2016
## Plot saved for cohort year 2021
Total Medicaid Spending per Capita
run_synthdid(data, "all_medicaid_spending_per_cap", cohorts)
## Error in cohort year 1992: dim(X) must have a positive length
## Plot saved for cohort year 1994
## Plot saved for cohort year 1995
## Plot saved for cohort year 1996
## Plot saved for cohort year 1997
## Plot saved for cohort year 1998
## Plot saved for cohort year 2006
## Plot saved for cohort year 2007
## Plot saved for cohort year 2011
## Plot saved for cohort year 2012
## Plot saved for cohort year 2014
## Plot saved for cohort year 2016
## Plot saved for cohort year 2021
Log(Total Medicaid Spending per Capita)
run_synthdid(data, "log_all_medicaid_spending_per_cap", cohorts)
## Error in cohort year 1992: dim(X) must have a positive length
## Plot saved for cohort year 1994
## Plot saved for cohort year 1995
## Plot saved for cohort year 1996
## Plot saved for cohort year 1997
## Plot saved for cohort year 1998
## Plot saved for cohort year 2006
## Plot saved for cohort year 2007
## Plot saved for cohort year 2011
## Plot saved for cohort year 2012
## Plot saved for cohort year 2014
## Plot saved for cohort year 2016
## Plot saved for cohort year 2021
TWFE Treatment Effect by Cohort (cohorts based off 5-10 year
intervals)
# Define 5 and 10 year interval cohorts based on treatment years
data_cohorts <- data %>%
mutate(cohort_5_interval = case_when(
treatment_year >= 1991 & treatment_year <= 1995 ~ "1991-1995",
treatment_year >= 1996 & treatment_year <= 2000 ~ "1996-2000",
treatment_year >= 2001 & treatment_year <= 2005 ~ "2001-2005",
treatment_year >= 2006 & treatment_year <= 2010 ~ "2006-2010",
treatment_year >= 2011 & treatment_year <= 2015 ~ "2011-2015",
treatment_year >= 2016 & treatment_year <= 2022 ~ "2016-2022",
TRUE ~ "never-treated"
)) %>%
mutate(cohort_10_interval = case_when(
treatment_year >= 1991 & treatment_year <= 2000 ~ "1991-2000",
treatment_year >= 2001 & treatment_year <= 2010 ~ "2001-2010",
treatment_year >= 2011 & treatment_year <= 2022 ~ "2011-2022",
TRUE ~ "never-treated"
))
# Vector of relative time variables
keepvars <- c(paste0("`center_time_-", 10:2, "`"),
paste0("center_time_", 0:10))
prep_es_10 <- function(mod){
mod_2 <- tibble(
estimate = mod$coefficients,
term1 = rownames(mod$coefficients),
se = mod$se,
pval = mod$pval
)
# Filter coefficients for relative time leads and lags
es <- mod_2 %>%
filter(term1 %in% keepvars) %>%
mutate(t = c(-10:-2, 0:10)) %>%
select(t, estimate, se, pval)
# Add a row for t = -1 (reference period)
es <- rbind(es, c(-1, 0, 0, 0))
# Add confidence intervals
es <- es %>%
mutate(lower = estimate - 1.96 * se,
upper = estimate + 1.96 * se)
return(es)
}
prep_es_trend_10 <- function(mod){
mod_2 <- tibble(
estimate = mod$coefficients,
term1 = rownames(mod$coefficients),
se = mod$se,
pval = mod$pval
)
es <- mod_2 %>%
filter(term1 %in% keepvars) %>%
mutate(t = c(0:10)) %>%
select(t, estimate, se, pval)
es <- rbind(es, c(-1, 0, 0, 0))
es <- es %>%
mutate(lower = estimate - 1.96 * se,
upper = estimate + 1.96 * se,
star = ifelse(pval <= 0.001, "***",
ifelse(pval<=0.01,"**",
ifelse(pval<=0.05, "*", ""))),
coef_star = paste0(round(estimate,2), star, sep =""))
return(es)
}
#' Plot 10-year event study plot with data filtered by interval cohort
#' @param data panel data
#' @param outcome outcome variable (numeric)
#' @param outcome_name name of the outcome variable (character)
#' @param cohort_var name of the interval variable (character)
#' @param cohort_interval range of years (cohort) we want to filter for (character)
did_graph_trend_10 <- function(data, outcome, outcome_name, cohort_var, cohort_interval){
p <- list()
cohort_data <- data %>%
filter(!!sym(cohort_var) == cohort_interval)
# Normal event study
formula_str1 <- paste0(
outcome, " ~ ",
"`center_time_-10` + `center_time_-9` + ",
"`center_time_-8` + `center_time_-7` + `center_time_-6` + `center_time_-5` + ",
"`center_time_-4` + `center_time_-3` + `center_time_-2` + `center_time_0` + ",
"`center_time_1` + `center_time_2` + `center_time_3` + `center_time_4` + ",
"`center_time_5` + `center_time_6` + `center_time_7` + `center_time_8` + ",
"`center_time_9` + `center_time_10` | ",
"as.factor(state) + as.character(year) ",
sep = ""
)
formula1 <- as.formula(formula_str1)
mod_d <- lfe::felm(formula1,
data = subset(cohort_data, center_time >= -10 & center_time <= 10),
exactDOF = TRUE)
es_d <- prep_es_10(mod_d)
es_d <- es_d %>%
mutate(sig = ifelse((lower > 0 | upper < 0), 1, 0))
# Pre-trend line
pre_trend <- es_d %>% filter(t < 0)
pre_trend_fit <- lm(estimate ~ t, data = pre_trend)
predicted_df <- data.frame(
t = seq(-10, 10, by = 0.1),
estimate = predict(pre_trend_fit, newdata = data.frame(t = seq(-10, 10, by = 0.1)))
)
## estimate difference
formula_str2 <- paste0(
outcome, " ~ ",
"`center_time_0` + ",
"`center_time_1` + `center_time_2` + `center_time_3` + `center_time_4` + ",
"`center_time_5` + `center_time_6` + `center_time_7` + `center_time_8` + ",
"`center_time_9` + `center_time_10` + ",
"center_time | ",
"as.factor(state) + as.character(year) ",
sep = ""
)
formula2 <- as.formula(formula_str2)
mod_d <- lfe::felm(formula2,
data = subset(cohort_data, center_time >= -10 & center_time <= 10),
exactDOF = TRUE)
es_d2 <- prep_es_trend_10(mod_d)
es_d2 <- es_d2 %>%
mutate(sig = ifelse((lower > 0 | upper < 0), 1, 0))
# Event study plot
p1 <- ggplot(data = es_d, aes(x = t, y = estimate, group = 1))+
geom_point(size= 3, color = "gray26")+
geom_errorbar(aes(ymin = lower, ymax = upper),
width=.1,
linewidth = 0.8,
color = "gray26") +
geom_hline(yintercept = 0,
linetype = "dashed",
linewidth = 1,
color = "firebrick4")+
scale_x_continuous(breaks = round(seq(-10, 10, by = 1),1),
limits = c(-10.5, 10.5))+
# scale_y_continuous(breaks = round(seq(-0.2, 0.2, by = 0.1), 1),
# limits = c(-0.2, 0.2))+
labs(x = "Relative Time", y = "Estimate",
title = paste0(outcome_name, " - ", cohort_interval))+
theme(plot.title = element_text(hjust = 0.5, size = 15),
legend.position = "none")+
geom_text(data = es_d2 %>% filter(t >= 0 & t <= 10),
aes(x = t, y = 0.5*max(es_d2$upper), label = coef_star, angle = 45),
color = "blue", size = 4, vjust = -0.3)+
geom_line(data = predicted_df, aes(x = t, y = estimate),
color = "blue", linetype = "dashed", linewidth = 1.2)
# annotate("text", x = 8, y = text_pos, label = comb_est, size = 5)
p2 <- ggplot(data = es_d2, aes(x = t, y = estimate)) +
geom_point(size = 3, color = "black") + # Black dots for estimates
geom_errorbar(aes(ymin = lower, ymax = upper),
width = 0.2,
linewidth = 0.8,
color = "black") + # Error bars
geom_line(color = "black", linewidth = 1) + # Line connecting estimates
geom_hline(yintercept = 0, linetype = "dashed", color = "firebrick4", linewidth = 1) + # Reference line at y = 0
scale_x_continuous(breaks = seq(-20, 20, by = 1)) +
# scale_y_continuous(breaks = seq(-0.5,1,0.1),
# limits = c(-0.5, 1.1))+
labs(x = "Relative Time",
y = "Estimate",
title = paste0(outcome_name, " - ", cohort_interval)) +
theme(plot.title = element_text(hjust = 0.5, size = 15),
legend.position="none")
p2
p[[1]] <- p1
p[[2]] <- p2
return(p)
}
# Define treatment cohorts
assign_cohort_5 <- function(treatment_year) {
case_when(
is.na(treatment_year) ~ "never-treated",
treatment_year >= 1996 & treatment_year <= 2000 ~ "1996-2000",
treatment_year >= 1991 & treatment_year <= 1995 ~ "1991-1995",
treatment_year >= 2006 & treatment_year <= 2010 ~ "2006-2010",
treatment_year >= 2016 & treatment_year <= 2022 ~ "2016-2022",
treatment_year >= 2011 & treatment_year <= 2015 ~ "2011-2015",
TRUE ~ "never-treated" # Default fallback if none of the above match
)
}
# Construct stacked data
# For each treatment year, construct control without treatment within 10/20 years
collect_10 <- list()
collect_20 <- list()
for(i in unique(data_cohorts$treatment_year[data_cohorts$treatment_year != Inf])){
# Get treated state
treated_state <- data_cohorts %>%
filter(treatment_year == i) %>%
mutate(treatment_year_group = i) %>%
mutate(cohort_5_interval = assign_cohort_5(treatment_year_group)) %>%
select(state, year, center_time, treated, total_medicaid,
log_all_medicaid_spending_per_cap, all_medicaid_spending_per_cap,
pct_in_comp_mco, treatment_year, treatment_year_group, cohort_5_interval)
max_year <- max(treated_state$year)
# Get control states within a 10 year (pre and post) from treatment year
control_state_10 <- data_cohorts %>%
filter(treatment_year > i + 10,
year <= i + 10, year >= i - 10) %>%
mutate(treatment_year_group = i,
center_time = 0,
treated = 0,
cohort_5_interval = assign_cohort_5(treatment_year_group)) %>%
select(state, year, center_time, treated, total_medicaid,
log_all_medicaid_spending_per_cap, all_medicaid_spending_per_cap,
pct_in_comp_mco, treatment_year, treatment_year_group, cohort_5_interval)
control_state_20 <- data_cohorts %>%
filter(treatment_year > i + 20,
year <= i + 20, year >= i - 20) %>%
mutate(treatment_year_group = i,
center_time = 0,
treated = 0,
cohort_5_interval = assign_cohort_5(treatment_year_group)) %>%
select(state, year, center_time, treated, total_medicaid,
log_all_medicaid_spending_per_cap, all_medicaid_spending_per_cap,
pct_in_comp_mco, treatment_year, treatment_year_group, cohort_5_interval)
collect_10[[length(collect_10)+1]] <- rbind(treated_state, control_state_10)
collect_20[[length(collect_20)+1]] <- rbind(treated_state, control_state_20)
}
# Collapse into one dataframe
dat_10 <- do.call(rbind, collect_10)
dat_20 <- do.call(rbind, collect_20)
dat_10 <- dat_10 %>% dummy_cols(select_columns = "center_time")
dat_20 <- dat_20 %>% dummy_cols(select_columns = "center_time")
# DID - 10 period
dat <- dat_10
dat <- dat %>%
mutate(log_total_spend = log(total_medicaid))
Plots for 5 year interval cohorts
cohorts <- c("1991-1995", "1996-2000", "2006-2010", "2011-2015", "2016-2022")
outcomes <- list(
pct_in_comp_mco = "Prop. Comp MCO Enroll",
log_total_spend = "Log(Total Medicaid Spending)",
log_all_medicaid_spending_per_cap = "Log(Total Medicaid Spending per Capita)"
)
# Initialize an empty list to store the plots
plot_list <- list()
# Loop through each outcome
for (outcome_name in names(outcomes)) {
outcome_label <- outcomes[[outcome_name]]
# Loop through each cohort
for (cohort in cohorts) {
# Generate the plot
p <- did_graph_trend_10(dat, outcome_name, outcome_label, "cohort_5_interval", cohort)
# Store the plot in the list
plot_list[[paste(outcome_name, cohort, sep = "_")]] <- p
}
outcome_plots <- plot_list[grep(outcome_name, names(plot_list))]
print(outcome_plots)
}
## $`pct_in_comp_mco_1991-1995`
## $`pct_in_comp_mco_1991-1995`[[1]]
##
## $`pct_in_comp_mco_1991-1995`[[2]]
##
##
## $`pct_in_comp_mco_1996-2000`
## $`pct_in_comp_mco_1996-2000`[[1]]
##
## $`pct_in_comp_mco_1996-2000`[[2]]
##
##
## $`pct_in_comp_mco_2006-2010`
## $`pct_in_comp_mco_2006-2010`[[1]]
##
## $`pct_in_comp_mco_2006-2010`[[2]]
##
##
## $`pct_in_comp_mco_2011-2015`
## $`pct_in_comp_mco_2011-2015`[[1]]
##
## $`pct_in_comp_mco_2011-2015`[[2]]
##
##
## $`pct_in_comp_mco_2016-2022`
## $`pct_in_comp_mco_2016-2022`[[1]]
##
## $`pct_in_comp_mco_2016-2022`[[2]]
##
##
## $`log_total_spend_1991-1995`
## $`log_total_spend_1991-1995`[[1]]
##
## $`log_total_spend_1991-1995`[[2]]
##
##
## $`log_total_spend_1996-2000`
## $`log_total_spend_1996-2000`[[1]]
##
## $`log_total_spend_1996-2000`[[2]]
##
##
## $`log_total_spend_2006-2010`
## $`log_total_spend_2006-2010`[[1]]
##
## $`log_total_spend_2006-2010`[[2]]
##
##
## $`log_total_spend_2011-2015`
## $`log_total_spend_2011-2015`[[1]]
##
## $`log_total_spend_2011-2015`[[2]]
##
##
## $`log_total_spend_2016-2022`
## $`log_total_spend_2016-2022`[[1]]
##
## $`log_total_spend_2016-2022`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_1991-1995`
## $`log_all_medicaid_spending_per_cap_1991-1995`[[1]]
##
## $`log_all_medicaid_spending_per_cap_1991-1995`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_1996-2000`
## $`log_all_medicaid_spending_per_cap_1996-2000`[[1]]
##
## $`log_all_medicaid_spending_per_cap_1996-2000`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_2006-2010`
## $`log_all_medicaid_spending_per_cap_2006-2010`[[1]]
##
## $`log_all_medicaid_spending_per_cap_2006-2010`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_2011-2015`
## $`log_all_medicaid_spending_per_cap_2011-2015`[[1]]
##
## $`log_all_medicaid_spending_per_cap_2011-2015`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_2016-2022`
## $`log_all_medicaid_spending_per_cap_2016-2022`[[1]]
##
## $`log_all_medicaid_spending_per_cap_2016-2022`[[2]]
# Define treatment cohorts
assign_cohort_10 <- function(treatment_year) {
case_when(
treatment_year >= 1991 & treatment_year <= 2000 ~ "1991-2000",
treatment_year >= 2001 & treatment_year <= 2010 ~ "2001-2010",
treatment_year >= 2011 & treatment_year <= 2022 ~ "2011-2022",
TRUE ~ "never-treated" # Default fallback if none of the above match
)
}
# Construct stacked data
# For each treatment year, construct control without treatment within 10/20 years
collect_10 <- list()
collect_20 <- list()
for(i in unique(data_cohorts$treatment_year[data_cohorts$treatment_year != Inf])){
# Get treated state
treated_state <- data_cohorts %>%
filter(treatment_year == i) %>%
mutate(treatment_year_group = i) %>%
mutate(cohort_10_interval = assign_cohort_10(treatment_year_group)) %>%
select(state, year, center_time, treated, total_medicaid,
log_all_medicaid_spending_per_cap, all_medicaid_spending_per_cap,
pct_in_comp_mco, treatment_year, treatment_year_group, cohort_10_interval)
max_year <- max(treated_state$year)
# Get control states within a 10 year (pre and post) from treatment year
control_state_10 <- data_cohorts %>%
filter(treatment_year > i + 10,
year <= i + 10, year >= i - 10) %>%
mutate(treatment_year_group = i,
center_time = 0,
treated = 0,
cohort_10_interval = assign_cohort_10(treatment_year_group)) %>%
select(state, year, center_time, treated, total_medicaid,
log_all_medicaid_spending_per_cap, all_medicaid_spending_per_cap,
pct_in_comp_mco, treatment_year, treatment_year_group, cohort_10_interval)
control_state_20 <- data_cohorts %>%
filter(treatment_year > i + 20,
year <= i + 20, year >= i - 20) %>%
mutate(treatment_year_group = i,
center_time = 0,
treated = 0,
cohort_10_interval = assign_cohort_10(treatment_year_group)) %>%
select(state, year, center_time, treated, total_medicaid,
log_all_medicaid_spending_per_cap, all_medicaid_spending_per_cap,
pct_in_comp_mco, treatment_year, treatment_year_group, cohort_10_interval)
collect_10[[length(collect_10)+1]] <- rbind(treated_state, control_state_10)
collect_20[[length(collect_20)+1]] <- rbind(treated_state, control_state_20)
}
dat_10 <- do.call(rbind, collect_10)
dat_20 <- do.call(rbind, collect_20)
dat_10 <- dat_10 %>% dummy_cols(select_columns = "center_time")
dat_20 <- dat_20 %>% dummy_cols(select_columns = "center_time")
# DID - 10 period
dat <- dat_10
dat <- dat %>%
mutate(log_total_spend = log(total_medicaid))
Plots for 10 year interval cohorts
cohorts2 <- c("1991-2000", "2001-2010", "2011-2022")
outcomes <- list(
pct_in_comp_mco = "Prop. Comp MCO Enroll",
log_total_spend = "Log(Total Medicaid Spending)",
log_all_medicaid_spending_per_cap = "Log(Total Medicaid Spending per Capita)"
)
# Initialize an empty list to store the plots
plot_list <- list()
# Loop through each outcome
for (outcome_name in names(outcomes)) {
outcome_label <- outcomes[[outcome_name]]
# Loop through each cohort
for (cohort in cohorts2) {
# Generate the plot
p <- did_graph_trend_10(dat, outcome_name, outcome_label, "cohort_10_interval", cohort)
# Store the plot in the list
plot_list[[paste(outcome_name, cohort, sep = "_")]] <- p
}
outcome_plots <- plot_list[grep(outcome_name, names(plot_list))]
print(outcome_plots)
}
## $`pct_in_comp_mco_1991-2000`
## $`pct_in_comp_mco_1991-2000`[[1]]
##
## $`pct_in_comp_mco_1991-2000`[[2]]
##
##
## $`pct_in_comp_mco_2001-2010`
## $`pct_in_comp_mco_2001-2010`[[1]]
##
## $`pct_in_comp_mco_2001-2010`[[2]]
##
##
## $`pct_in_comp_mco_2011-2022`
## $`pct_in_comp_mco_2011-2022`[[1]]
##
## $`pct_in_comp_mco_2011-2022`[[2]]
##
##
## $`log_total_spend_1991-2000`
## $`log_total_spend_1991-2000`[[1]]
##
## $`log_total_spend_1991-2000`[[2]]
##
##
## $`log_total_spend_2001-2010`
## $`log_total_spend_2001-2010`[[1]]
##
## $`log_total_spend_2001-2010`[[2]]
##
##
## $`log_total_spend_2011-2022`
## $`log_total_spend_2011-2022`[[1]]
##
## $`log_total_spend_2011-2022`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_1991-2000`
## $`log_all_medicaid_spending_per_cap_1991-2000`[[1]]
##
## $`log_all_medicaid_spending_per_cap_1991-2000`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_2001-2010`
## $`log_all_medicaid_spending_per_cap_2001-2010`[[1]]
##
## $`log_all_medicaid_spending_per_cap_2001-2010`[[2]]
##
##
## $`log_all_medicaid_spending_per_cap_2011-2022`
## $`log_all_medicaid_spending_per_cap_2011-2022`[[1]]
##
## $`log_all_medicaid_spending_per_cap_2011-2022`[[2]]
De Chaisemartin and D’Haultfoeuille (2020) (Callaway-Sant’anna)
Percent in Comprehensive MCO
Total Medicaid Enrollment
Total Medicaid Spending
Log(Total Medicaid Spending)
Total Medicaid Spending per Capita
Log(Total Medicaid Spending per Capita)
